[azizi0407]

I'm simulating a natural convection in fluent using bossinesq approximation. I was told that the temperature gradient of the system cannot be large. So, what is the range of the tempareture gradient that acceptable to use in fluent?

In experimental works, Rayleigh number of the natural convection is varies by the change of the temperature gradient. In CFD, what variable can we change to increased the Rayleigh number as the temperature gradient cannot be increase too much?

[LuckyTran]

These are not CFD limitations. There are no limitations in Fluent, you impose the Boussinesq approximation and that's it.

The boussinesq approximation poses that the density change is linearly proportional to the temperature change. Boussinesq is valid as long as this is true physically.

If temperature variations are large then the Boussinesq approximation is no a longer valid physical approximation. Again, this is fault of the numerics but a modeling error and user fault.

Use an explicit equation of state like the ideal gas law if you don't want to deal with the temperature range issue.

[azizi0407]

Thank you for your answer.

[Kader]

Originally Posted by LuckyTran

These are not CFD limitations. There are no limitations in Fluent, you impose the Boussinesq approximation and that's it.

The boussinesq approximation poses that the density change is linearly proportional to the temperature change. Boussinesq is valid as long as this is true physically.

If temperature variations are large then the Boussinesq approximation is no a longer valid physical approximation. Again, this is fault of the numerics but a modeling error and user fault.

Use an explicit equation of state like the ideal gas law if you don't want to deal with the temperature range issue.

The limitation is about Deata T=27°